GitLab

修改测试用例

* Add bf16 template sp

* prepare merge

* add enable bf

* add bf16 decode attention support

* fix python lint

* fix yapf

* fix c format

* c format11

* fix cast

* fix on sm<80

* fix linux bf162 cast

* fix type cast

* fix lint

* support from hf pretrained

* fix pybind

* fix converter

* add trust remote code

* fix comment

* fix convert qwen

* fix lint

* fix baichuan

* update weight map

* support image_embs input

* add some checks

* update interactive/config.pbtxt && TurbomindModelConfig

* update docstring

* refactor

* support convert embeddings to bf16

* update interactive/config.pbtxt

* embeddings -> input_embeddings

* use input_embedding_ranges

* remove embedding_begins/ends

* disable attention mask when not needed

* fix for sm<80 and float data type

* simplify block manager

* fix lint

* Unify prefill and decode passes

* dynamic split-fuse

* refactor

* correct input count calculation

* remove unused

* lint

* lint

* fix msvc build

* fix msvc build

* fix msvc build

* fix msvc build

* fix msvc build

* fix msvc build

* fix msvc build

* fix msvc build

* fix msvc build

* feat(build): enable ninja and lld

* fix(.github): add ninja installation

* fix(CI): remove dimsize=256

* fix(CI): add option for generate.sh

* fix(docs): update

* fix

* fix lint

* turbomind support export model params

* fix overflow

* support turbomind.from_pretrained

* fix tp

* support AutoModel

* support load kv qparams

* update auto_awq

* udpate docstring

* export lmdeploy version

* update doc

* remove download_hf_repo

* LmdeployForCausalLM -> LmdeployForCausalLM

* refactor turbomind.py

* update comment

* add bfloat16 convert back

* support gradio run_locl load hf

* support resuful api server load hf

* add docs

* support loading previous quantized model

* adapt pr 690

* udpate docs

* not export turbomind config when quantize a model

* check model_name when can not get it from config.json

* update readme

* remove model_name in auto_awq

* update

* update

* udpate

* fix build

* absolute import

* tmp

* update

* update

* optimize for throughput

* update

* fix eos

* clean up

* fix serving

* fix indexed copy

* minor

* minor

---------
Co-authored-by: lvhan028 <lvhan_028@163.com>

* fix init of finished buf

* fix `finished_count`

* refresh decoder attention kernel

* block-level kv cache

* `BlockManager` & `SequenceManager`

* update

* update

* update

* update

* rename

* GQA support

* fix context length

* GQA dispatch

* kv8

* tune

* async stream cb

* nvtx

* config parsing

* debug

* optimize output cost

* split-k decoding

* minor

* truncate `session_len` by available blocks

* minor

* license

* fix

* dispatch `cp.async`

* fix linking

* fix

* fix deadlock

* guard input length

* correct start offset

* fix prefill chunking

* fix `cache_block_seq_len` param passing

* fix `block_size` fmtstr

* fix output tokens

* fix batch resizing

* fix masking of finished sequences

* add debug util

* free unused block early

* add ntk scaling and logn scaling

* cmake flags

* fix typo

* w4a16 for sm75

* fix msvc build

* fix msvc build

* fix block verification

* fix msvc build

* use `std::shuffle`

* fix lint

* fix lint

* fix lint

* clear incoming buffer

* clear finished requests

* fix batch initialization

* fix typo

* fix typo

* fix comparison

* change shared_instances_ from weakptr to sharedptr

* update

* Fix memory leak

* modern c++

* cuda allocator fix

* graceful termination

* lint and compilation fix

* support actual seqlen

* fix lint

* update variable types

* lint

* update type

* fix lint

---------

* reduce gil switching

* ffi lock func

* remove unused

* remove unused

* remove unused

* tmp

* add demo for codellama inference

* update

* update

* update

* update codellama.md

* export rope_theta

* update

* update doc

* fix client.py

* define SamplingParam

* rollback 'end'

* rotary_emb_base to rotary_embedding_base

* change to baichuan2-7b

* Fix crash when context window size is large by setting max dynamic smem size

* fix linting

* pack llama_gemm

* update CMakeLists.txt

* remove candidate

* update MANIFEST.in

* first

* fix causal mask

* disable flash attention2 on sm70

* fix 2

* update readme

* clang-format

* disable ft2 on windows

* fix lint

* fix build

* fix build

* fix long kv seq

* fix lint

* sync copy output

---------
Co-authored-by: grimoire <yaoqian@pjlab.org.cn>
Co-authored-by: irexyc <irexyc@gmail.com>

* Pad tok_embedding and output weights to make their shape divisible by TP

* update

* update

* update

* update

* update llamaBatch

* disable cache hint for CUDA < 11.4

* fix lint

* fix lint

* fix cuda-11.3 build